CN112130148B - Land type-based DS self-adaptive selection method in InSAR time sequence analysis - Google Patents

Land type-based DS self-adaptive selection method in InSAR time sequence analysis Download PDF

Info

Publication number
CN112130148B
CN112130148B CN202010960574.9A CN202010960574A CN112130148B CN 112130148 B CN112130148 B CN 112130148B CN 202010960574 A CN202010960574 A CN 202010960574A CN 112130148 B CN112130148 B CN 112130148B
Authority
CN
China
Prior art keywords
land
homogeneous
time sequence
pixels
pixel
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010960574.9A
Other languages
Chinese (zh)
Other versions
CN112130148A (en
Inventor
曾琪明
张志亮
焦健
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Peking University
Original Assignee
Peking University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Peking University filed Critical Peking University
Priority to CN202010960574.9A priority Critical patent/CN112130148B/en
Publication of CN112130148A publication Critical patent/CN112130148A/en
Application granted granted Critical
Publication of CN112130148B publication Critical patent/CN112130148B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • G01S13/89—Radar or analogous systems specially adapted for specific applications for mapping or imaging
    • G01S13/90—Radar or analogous systems specially adapted for specific applications for mapping or imaging using synthetic aperture techniques, e.g. synthetic aperture radar [SAR] techniques
    • G01S13/9021—SAR image post-processing techniques
    • G01S13/9023—SAR image post-processing techniques combined with interferometric techniques

Landscapes

  • Engineering & Computer Science (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Radar Systems Or Details Thereof (AREA)
  • Image Analysis (AREA)

Abstract

The invention discloses a DS self-adaptive selection method in InSAR time sequence analysis based on land types, wherein in the DS self-adaptive selection process in the InSAR time sequence analysis, the identification of homogeneous pixels on a time sequence SAR image is carried out based on the land types, and then DS candidate points are selected; the land type is a land covering type or a land utilization type; then, adaptively calculating DS selection thresholds under different land types according to different land types; and then DS self-adaptive selection based on the land type is carried out. By adopting the technical scheme of the self-adaptive selection for the DS target by InSAR time sequence analysis based on the land type, the quality of the self-adaptive selected DS can be improved; the inspection search range is narrowed, the recognition efficiency of the homogeneous pixels is improved, and the DS selection efficiency is further improved.

Description

Land type-based DS self-adaptive selection method in InSAR time sequence analysis
Technical Field
The invention relates to a method for selecting a distributed scatterer in Interferometric Synthetic Aperture Radar (InSAR) time sequence analysis, in particular to a method for adaptively selecting a Distributed Scatterer (DS) based on a land cover/land utilization type (land type).
Background
The InSAR time sequence analysis technology is an important means for monitoring the slow deformation of the earth surface. The method comprises the steps of analyzing a point-shaped Permanent Scatterer (PS) or a planar Distributed Scatterer (DS) on a time series SAR image to obtain the surface deformation information of a monitored area on the time series. The PS and the DS are collectively called deformation measurement points, and the quantity and the quality of the deformation measurement points directly influence the precision and the reliability of the InSAR time sequence analysis result.
The initial research object of the InSAR timing analysis technology is the PS target on the timing SAR image, which mainly corresponds to artificial ground objects such as houses, bridges, dams and the like in the field or huge bare rocks facing the incident direction of radar waves, and since many artificial buildings and strong reflection targets exist in urban areas, the technology has been mainly applied to monitoring surface deformation caused by groundwater extraction, underground engineering construction and the like in urban areas in the past, and the appearance of the timing InSAR technology considering DS has changed the situation. The DS generally corresponds to the land object types such as bare land, desert and the like with low vegetation coverage, and increases the number of deformation measuring points in the original InSAR time sequence analysis, so that the technology can be expanded to non-urban area ground surface deformation monitoring caused by volcanic activity, earthquake, landslide and the like. However, currently, a single selection threshold is generally adopted in the DS selection process by using the DS InSAR timing analysis technology, and as a result, distribution of DS in a research area is extremely unbalanced, and an extreme case that DS in a large area is rare or there is no DS occurs, so that a timing InSAR resolving result is unreliable or even a case that there is no solution. The selection threshold of the DS is mechanically reduced, the number of the DS can be increased to a certain extent, but some DSs with poor quality can be selected, the accuracy of the time sequence InSAR calculation result can be influenced, and the calculation time is increased. In addition, the core step of DS selection is to identify spatially adjacent pixels that obey the same statistical distribution on the sequential SAR image, i.e., the identification of homogeneous pixels, which is time-consuming in the processing process and prevents the popularization and application of the technique in the field of surface deformation monitoring to a certain extent.
In the prior art, an Adaptive DS-InSAR method proposed by Jia et al (Jia H, Zhang H, Liu L, et al. landslide development Monitoring by Adaptive Distributed scattering [ J ] removal Sensing,2019,11(19)) divides a research area uniformly and regularly in sequence according to N levels (e.g. 8, 16, 32, etc.), balances the selection number and time of DS under different N values, and finally determines the value of N to realize the partitioned Adaptive selection of DS, which can improve the spatial distribution density of DS in a Deformation Monitoring area, but adopts a mechanical regular geometric partitioning mode, does not consider the natural attributes or the factors of the state of DS, reduces the correct rate of the selection of DS, thereby affecting the timing result of the selected DS and the accuracy of the selected DS in the sar process of repeated sar test, the time cost and computational burden of the solution are increased.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides a novel method for adaptively selecting a Distributed Scatterer (DS) based on a land cover/land utilization type, which takes a land cover type graph or a land utilization type graph as a constraint condition for selecting the DS in a deformation monitoring area, adaptively calculates a selection threshold of the DS aiming at different land cover/land utilization types (hereinafter, the land cover/land utilization types are simply referred to as land types), and finally realizes the DS adaptive selection based on the land types.
The technical scheme provided by the invention is as follows:
a DS self-adaptive selection method in InSAR time sequence analysis based on land type, in the process of DS self-adaptive selection in InSAR time sequence analysis, the identification of homogeneous pixels on a time sequence SAR image is carried out based on land coverage/land utilization type (land type), and then DS Candidate points (Distributed scanner Candidate points, DSC) are selected; then, adaptively calculating DS selection thresholds under different land types according to different land types; then DS self-adaptive selection based on the land type is carried out; the method comprises the following steps:
s1) land type identification of SAR image pixels under a radar coordinate system is carried out:
and introducing a land type graph, converting the land type graph under a geographic coordinate system into a radar coordinate system of the SAR image, and identifying the land type of the SAR image pixel under the radar coordinate system. The method comprises the following steps:
s11) carrying out geometric registration on the N time sequence SAR images to obtain a registered time sequence SAR image; and establishing a corresponding relation between SAR image pixel radar coordinates and geographic coordinates.
S12) determining the pixel position corresponding to the geographic coordinate of the SAR image pixel on the land type map by adopting a nearest neighbor method according to the geographic coordinate of the SAR image pixel in S11), thereby establishing the pixel corresponding relation between the land type map under the geographic coordinate system and the SAR image under the radar coordinate system, converting the land type map under the geographic coordinate system into the radar coordinate system, and finding the land type attribute value (namely, type value) corresponding to the SAR image pixel to realize the land type identification of the SAR image pixel.
S2) selecting DS candidate points based on the land types: and identifying the homogeneous pixel on the time sequence SAR image based on the land type, and then selecting the DS candidate points. The method comprises the following steps:
s21) determining a homogeneous pixel detection window by taking the land type as a constraint condition for homogeneous pixel identification, and identifying pixels on the sequential SAR image, which are homogeneous with pixels in the center of the window, by using goodness-of-fit inspection;
in the process of identifying the homogeneous pixel on the time sequence SAR image by using goodness-of-fit test (taking Kolmogorov-Smirnov test, namely K-S test as an example in the invention), a land type is introduced and used as a constraint condition for identifying the homogeneous pixel. Before K-S detection, firstly determining the size m x n of a homogeneous pixel detection window, wherein m is the pixel number in the window orientation direction, and n is the pixel number in the distance direction; then judging the land type of the central pixel of the homogeneous pixel detection window, and if the land type of the central pixel is water or is different from the land types of other pixels in the window, not performing K-S inspection on the central pixel; otherwise, performing K-S inspection to determine homogeneous pixels.
S22) performing connectivity detection on the central pixels and the homogeneous pixels in the homogeneous pixel detection window to obtain the number of the homogeneous pixels in the homogeneous pixel detection window;
and after the K-S inspection of the central pixel and all other pixels in the window is completed, performing connectivity detection on the central pixel and the homogeneous pixels thereof in the window, counting the number of pixels which are homogeneous and communicated with the central pixel, and adding the central pixel of the window to serve as the number of the homogeneous pixels in the window.
S23), setting a selection threshold of the DS candidate points, and selecting window center pixels as the DS candidate points according to the number of the homogeneous pixels in the homogeneous pixel detection window in the step S22)
The selection threshold of the DS candidate point is determined in advance, and the size of the selection threshold varies with the pixel size, the monitoring area, the size of the homogeneous pixel detection window, and other factors, and is determined empirically, but it must be a positive integer and cannot be greater than the total number of pixels of the homogeneous pixel detection window, that is, the value of the selection threshold is not greater than m × n. And when the number of the homogeneous pixels in the window is greater than or equal to the selection threshold of the DS candidate point, selecting the central pixel as the DS candidate point.
S3) setting an InSAR space-time baseline threshold, carrying out time sequence SAR image differential interference processing, and realizing land type-based self-adaptive determination of DS selection threshold through coherence;
self-adaptive determination of DS selection threshold based on land type: the DS selection threshold in the invention is essentially determined by coherence, and the coherence threshold selected by the DS under different land types is obtained by self-adaptively calculating according to the land types.
Before adaptively determining a DS selection threshold, firstly setting an InSAR space-time baseline threshold, and screening M image interference pairs from N time sequence SAR images, wherein the space-time baseline threshold is different due to different specific SAR parameters and is empirically determined, for example, for ALOS-2 satellite strip mode SAR data, the general time baseline threshold is set to be 250-400 days, and the space baseline threshold is set to be 500-1500M; for the Sentinel-1 satellite interference wide-amplitude mode SAR data, the general time baseline threshold is set to be 60-200 days, and the space baseline threshold is set to be 100-250 m. Then, the images are subjected to differential interference processing to generate M differential interference images. On the basis, the DS candidate points are taken as centers, the size of a DS candidate point coherence calculation window is set, and the size of the DS candidate point coherence calculation window is the same as that of a homogeneous pixel detection window and is m x n; then, the DS selection threshold values under different land types are determined adaptively through the following steps:
s31) calculating the coherence of each DS candidate point on each differential interferogram of the time sequence SAR image:
the coherence of each DS candidate point on each differential interferogram is calculated by the following formula (1):
Figure BDA0002680404290000041
wherein, cork(w) is the w-th DS candidate point (w 1,2, … SUM) on the k-th differential interferogram (k 1,2DSC) Coherency, SUM ofDSCThe total number of DS candidate points in the whole deformation monitoring area; sk,1(. and s)k,2(·) respectively representing the master and slave images of the SAR used to generate the kth differential interferogram, the superscript denotes the complex conjugate; num (w) and l are the number of homogeneous pixels in the window and the number of homogeneous pixels, respectively.
S32) calculating the time sequence average of the DS candidate point coherence:
the time sequence average value of the coherence of each DS candidate point is calculated by the following formula (2):
Figure BDA0002680404290000042
wherein, w is the serial number of the DS candidate point;
Figure BDA0002680404290000043
represents the coherence timing average of the w-th DS candidate point, and M is the total number of differential interferograms.
S33) calculating a spatial mean of the coherence time-series averages of the DS candidate points:
a) calculating the average spatial mean value of the coherence time sequence of DS candidate points in the whole deformation monitoring area, and calculating by the formula (3):
Figure BDA0002680404290000044
wherein the content of the first and second substances,
Figure BDA0002680404290000051
spatial mean, SUM, representing the mean of the coherence time sequence of DS candidate points over the entire deformation monitoring areaDSCThe total number of DS candidate points in the whole deformation monitoring area.
b) And (4) calculating the spatial mean value of the DS candidate point coherence time sequence average under different land types, and calculating by using the formula (4).
Figure BDA0002680404290000052
Wherein the content of the first and second substances,
Figure BDA0002680404290000053
a spatial mean value representing a coherence time-series average of the DS candidate points under an I-th land type, I being an identification code of the land type (I ═ 1, 2., total number of land types);
Figure BDA0002680404290000054
is the total number of DS candidate points under the I-th land type,
Figure BDA0002680404290000055
the number of DS candidate points under different land types is respectively counted according to the land type corresponding to each DS candidate point.
S34) calculating DS selection thresholds for each land type:
calculated by equation (5):
Figure BDA0002680404290000056
wherein, corthreshold(I) Denotes the DS coherence picking threshold for type I land, min (-) denotes the lesser of the two values in parentheses.
S4) self-adaptive selection of the DS is carried out; the method comprises the following steps:
s41) according to the land type I to which the DS candidate point w belongs, averaging the coherence time sequence of the DS candidate point
Figure BDA0002680404290000057
Threshold cor is selected according to DS coherence under the land type to which the threshold cor belongsthreshold(I) Make a comparison if
Figure BDA0002680404290000058
Selecting the DS candidate point as the DS; otherwise, the point is not selected.
S42) repeating step S41) until all DS candidate points have been processed; obtaining the final result of the DS self-adaptive selection;
through the steps, the DS is selected in an InSAR time sequence analysis based on the land type in a self-adaptive mode.
Compared with the prior art, the invention has the beneficial effects that:
in the self-adaptive selection of the InSAR time sequence analysis DS, a land type graph is introduced, so that an SAR image pixel carries land type information, the land type can be used as a constraint condition for judging a homogeneous pixel in the initial selection stage of the DS, and the situation that the pixel with the land type of a water body or the pixel which does not belong to the same type as the pixel at the center of a detection window participates in K-S detection can be avoided, so that the search range of the K-S detection can be reduced, and the identification efficiency of the homogeneous pixel is improved.
And secondly, the DS selection threshold value can be determined and the DS can be selected in a self-adaptive manner aiming at different land types by introducing the land type diagram, so that the DS selection threshold value selecting method can improve the selection quality of the DS by considering the natural attribute or the utilization state characteristic factor of the DS. In addition, unlike the existing adaptive DS-InSAR method, the present invention does not require regular geometric partitioning, thus avoiding the subjectivity of partitioning regions and the time it takes to repeat the test to determine the region level.
Drawings
FIG. 1 is a flow chart of a DS adaptive selection method based on land type provided by the invention.
FIG. 2 is a block diagram of a DS candidate selection process in an embodiment of the present invention.
FIG. 3 is a block diagram of a flow chart of adaptive calculation of DS selection threshold values under different land types in an embodiment of the present invention.
Fig. 4 is a block diagram of a flow of adaptive selection of a DS in an embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings.
The invention provides a novel method for adaptively selecting distributed scatterers based on land cover/land use types (hereinafter referred to as land types), which takes a land cover type graph or a land use type graph (land type graph) as a constraint condition for selecting Distributed Scatterers (DS) in a deformation monitoring area, automatically calculates selection thresholds of the DSs under different land types aiming at different land types, and realizes the adaptive selection of the DSs.
Fig. 1 shows a flow of a DS adaptive selection method based on land type provided by the present invention, which specifically includes the following steps:
1) land type identification of SAR image pixel under radar coordinate system
Converting a land type map under a geographic coordinate system into a radar coordinate system of the SAR image, and identifying the land type of the SAR image pixel under the radar coordinate system, wherein the process is briefly described as follows:
1a) carrying out geometric registration on the N time sequence SAR images to obtain a registered time sequence SAR image; a lookup table between an SAR image radar coordinate system and a geographic coordinate system is established, and a corresponding relation between a pixel radar coordinate and a geographic coordinate is established;
1b) according to the geographic coordinates of the SAR image pixels in the radar coordinate system in the lookup table, further determining the positions of the pixels corresponding to the SAR image pixels in the geographic coordinate system on the land type graph by adopting a nearest neighbor method, and thus finding the land type attribute value (namely type value) corresponding to the pixels;
1c) and repeating the step 1b) until all the pixels of the SAR image have corresponding land type attribute values, obtaining a land type graph under a radar coordinate system, and finishing the land type identification of the SAR image pixels under the radar coordinate system.
2) SAR image differential interference processing
The method comprises the following steps of carrying out interference pair construction and differential interference processing on a time sequence SAR image, and finally generating a differential interference image, wherein the processing process is briefly described as follows:
2a) and setting an InSAR space-time baseline threshold, and screening M image interference pairs from the N time sequence SAR images. The spatiotemporal baseline threshold value is different due to different specific SAR working parameters and needs to be determined empirically. Generally, for the ALOS-2 satellite strip mode SAR data, the general time baseline threshold is set to 250-; for the Sentinel-1 satellite interference wide-amplitude mode SAR data, the general time baseline threshold is set to be 60-200 days, and the space baseline threshold is set to be 100-250 m.
2b) And respectively carrying out differential interference processing on the interference pair images to generate M differential interference images.
3) DSC selection on time sequential SAR images
Through goodness-of-fit inspection (taking K-S inspection as an example in the invention), pixels which are homogeneous with the central pixel in the detection window of each homogeneous pixel on the registered time sequence SAR image are determined, homogeneous pixel connectivity detection is carried out, and a DSC is screened according to a DSC selection threshold, wherein the specific implementation flow is shown in figure 2.
3a) Setting the size of the homogeneous pixel detection window as m × n, wherein m is the pixel number in the window orientation direction, and n is the pixel number in the distance direction.
3b) Firstly, judging whether the land type of the center pixel of the window is a water body. If the pixel is a water body, the pixel is not subjected to any subsequent treatment; if the water body is not the water body, continuing to execute the step 3 c).
3c) And for the non-water body, performing pixel homogeneity inspection. Firstly, comparing the type values of the central pixel of the window with the type values of other pixels in the window, and if the type values of the central pixel and the type values of other pixels in the window are not equal, namely the central pixel and the type values belong to different land types, not performing subsequent treatment. If the type values of the two are equal, namely the two belong to the same land type, performing K-S inspection on the two to judge whether the two belong to the same pixel, and if the judgment result is that the two are the same, marking the non-central pixel as 1; if the two are not homogeneous, the non-center pixel element is marked as 0.
3d) And repeating the step 3c) until the central pixel and all other pixels in the window are subjected to homogeneity test. Then, carrying out connectivity detection on the central pixel and the homogeneous pixel thereof in the window; and (4) counting the number of pixels which are homogeneous and communicated with the central pixel, adding the central pixel as the number of homogeneous pixels of the window, and continuously executing the step 3 e).
3e) The DSC selection threshold is determined in advance, and the size of the DSC selection threshold is different due to factors such as the size of pixels, a monitoring area, the size of a homogeneous pixel detection window and the like, so that the DSC selection threshold is determined by experience, but the DSC selection threshold is a positive integer and cannot be larger than the total number of pixels of the homogeneous pixel detection window, namely the value of the DSC selection threshold is not larger than m x n. And (3) comparing the number of the homogeneous pixels in the window with a DSC selection threshold, if the number of the homogeneous pixels is less than the DSC selection threshold, moving the central pixel to the next position, performing new window processing, and restarting to execute the step 3 b). And if the number of the homogeneous pixels is greater than or equal to the DSC selection threshold, selecting the central pixel of the window as the DSC.
3f) And 3b) repeating the process until all internal pixels of the time sequence SAR image are traversed as central pixels of the homogeneous pixel detection window, and finally obtaining the DSC selection result in the whole deformation monitoring area. Here, the intra pixels are pixels excluding the peripheral pixels (whose width is 1/2 the size of the detection window in the corresponding direction), and the total number of the pixels is defined as totalin-pixel。
4) Adaptive calculation of DS selection threshold
And sequentially carrying out coherence calculation on all DSCs on each differential interference pattern, carrying out average calculation on a time sequence, and respectively carrying out space average calculation on the differential interference patterns in the whole deformation monitoring area and under different land types. On this basis, DS selection thresholds under corresponding types are calculated for different land types, and the implementation flow is shown in fig. 3, and the process is briefly described as follows:
4a) and setting the size of a DSC coherence calculation window by taking DSC as the center, wherein the size of the DSC coherence calculation window is the same as that of a homogeneous pixel detection window, and the size of the DSC coherence calculation window is m x n.
4b) Performing coherence calculation on all DSCs on each differential interference image in sequence, wherein only a central pixel (namely the DSC) and pixels which are homogeneous and communicated with the central pixel participate in calculation in a calculation window, and a calculation formula is shown as a formula (1):
Figure BDA0002680404290000091
wherein, cork(w) is the w-th DSC (w 1,2, … SUM) on the k-th differential interferogram (k 1,2DSC) Coherency, SUM ofDSCIs the total number of DSC in the whole deformation monitoring area; sk,1(. and s)k,2(. cndot.) represents the SAR master and slave images used to generate the kth differential interferogram, respectively, the superscript x representing the complex conjugate; num (w) and l refer to the number of homogeneous pixels in the window and the number of homogeneous pixels, respectively.
4c) The coherence of each DSC is respectively calculated by averaging in time sequence, and the calculation formula is shown as formula (2):
Figure BDA0002680404290000092
wherein the content of the first and second substances,
Figure BDA0002680404290000093
represents the coherence time-series average of the w-th DSC.
4d) Carrying out average calculation on the DSC coherence time sequence average value in the whole deformation monitoring area, wherein the calculation formula is shown as formula (3):
Figure BDA0002680404290000094
wherein the content of the first and second substances,
Figure BDA0002680404290000095
and (3) representing the spatial mean of the time sequence average value of DSC coherence in the whole deformation monitoring area.
4e) According to the land type corresponding to each DSC, counting the number of the DSCs in different land types; then, the DSC coherence time sequence average value is subjected to average value calculation under each land type, and the calculation formula is shown as formula (4):
Figure BDA0002680404290000096
wherein the content of the first and second substances,
Figure BDA0002680404290000097
a spatial mean value representing a DSC coherence time series average under the I-th land type, I being an identification code of the land type (I ═ 1, 2., total number of land types),
Figure BDA0002680404290000101
is the total number of DSC in type I land.
4f) Respectively calculating DS selection threshold values under each land type, wherein the calculation formula is shown as the formula (5):
Figure BDA0002680404290000102
wherein, corthreshold(I) Denotes the DS coherence picking threshold for type I land, min (-) denotes the lesser of the two values in parentheses.
5) DS adaptive selection
The specific implementation flow is shown in fig. 4:
5a) according to the land type (I) to which each DSC belongs, averaging the coherence time sequence of the DSC
Figure BDA0002680404290000103
Threshold cor is selected according to DS coherence under the land type to which the threshold cor belongsthreshold(I) Make a comparison if
Figure BDA0002680404290000104
Selecting the DSC as DS; otherwise, the DSC is not selected.
5b) And repeating the step 5a) until all the DSCs participate in the processing, and outputting the final result of the DS self-adaptive selection.
It is noted that the disclosed embodiments are intended to aid in further understanding of the invention, but those skilled in the art will appreciate that: various substitutions and modifications are possible without departing from the spirit and scope of the invention and appended claims. Therefore, the invention should not be limited to the embodiments disclosed, but the scope of the invention is defined by the appended claims.

Claims (7)

1. A DS self-adaptive selection method in InSAR time sequence analysis based on land types is characterized in that in the DS self-adaptive selection process in InSAR time sequence analysis, the identification of a homogeneous pixel on a time sequence SAR image is carried out based on the land types, and then DS candidate points are selected; the land type is a land covering type or a land utilization type; then, adaptively calculating DS selection thresholds under different land types according to different land types; then DS self-adaptive selection based on the land type is carried out; the method comprises the following steps:
s1) land type identification of SAR image pixels under a radar coordinate system is carried out:
converting the land type map in the geographic coordinate system into a radar coordinate system of the SAR image, and identifying the land type of the SAR image pixel in the radar coordinate system;
s2) selecting DS candidate points based on the land types:
identifying homogeneous pixels on the time sequence SAR image based on the land type, and then selecting DS candidate points; the method comprises the following steps:
s21) determining a homogeneous pixel detection window by taking the land type as a constraint condition for homogeneous pixel identification, and identifying pixels on the sequential SAR image, which are homogeneous with pixels in the center of the window, by using goodness-of-fit inspection;
s22) performing connectivity detection on the central pixels and the homogeneous pixels in the homogeneous pixel detection window to obtain the number of the homogeneous pixels in the homogeneous pixel detection window;
s23) setting a selection threshold of the DS candidate points, and selecting window center pixels as DS candidate points according to the number of the homogeneous pixels in the homogeneous pixel detection window in the step S22);
s3) setting an InSAR space-time baseline threshold, carrying out time sequence SAR image differential interference processing, and adaptively calculating and determining a DS selection threshold based on land types through coherence; the method comprises the following steps:
s31) calculating the coherence of each DS candidate point on each differential interferogram of the time sequence SAR image according to the formula (1):
Figure FDA0002680404280000011
wherein, cork(w) is the w-th DS candidate point (w 1,2, … SUM) on the k-th differential interferogram (k 1,2DSC) Coherency, SUM ofDSCThe total number of DS candidate points in the whole deformation monitoring area; sk,1() And sk,2() Respectively representing a SAR main image and a SAR slave image used for generating a k-th differential interferogram, and the upper mark indicates complex conjugate; num (w) and l are the number of homogeneous pixels in the window, respectivelyQuantity and number of homogeneous pixels;
s32) calculating the time-series average of the DS candidate point coherence by equation (2):
Figure FDA0002680404280000012
wherein, w is the serial number of the DS candidate point;
Figure FDA0002680404280000013
representing the coherence time sequence average value of the w-th DS candidate point, wherein M is the total number of the differential interferograms;
s33) calculating a spatial mean of the coherence time-series averages of the DS candidate points:
a) and (3) calculating to obtain a spatial mean value of the coherence time sequence average of the DS candidate points in the whole deformation monitoring area by the following formula:
Figure FDA0002680404280000021
wherein the content of the first and second substances,
Figure FDA0002680404280000022
spatial mean, SUM, representing the mean of the coherence time sequence of DS candidate points over the entire deformation monitoring areaDSCThe total number of DS candidate points in the whole deformation monitoring area is obtained;
b) calculating to obtain a spatial mean value of the DS candidate point coherence time sequence average under different land types through an equation (4);
Figure FDA0002680404280000023
wherein the content of the first and second substances,
Figure FDA0002680404280000024
spatial mean value representing the coherence time-series average of DS candidate points under the I land type, I being the identity of the land typeCode (I ═ 1,2, > total number of land types);
Figure FDA0002680404280000025
is the total number of DS candidate points under the I land type;
Figure FDA0002680404280000026
respectively counting the number of DS candidate points under different land types according to the land type corresponding to each DS candidate point;
s34) calculating the DS selection threshold value under each land type according to the formula (5):
Figure FDA0002680404280000027
wherein, corthreshold(I) The DS coherence selection threshold value under the I-th land type is represented, and min () represents the smaller of two numerical values in brackets;
s4) self-adaptive selection of the DS is carried out; the method comprises the following steps:
s41) according to the land type I to which the DS candidate point w belongs, averaging the coherence time sequence of the DS candidate point
Figure FDA0002680404280000028
Threshold cor is selected according to DS coherence under the land type to which the threshold cor belongsthreshold(I) Make a comparison if
Figure FDA0002680404280000029
Selecting the DS candidate point as the DS;
s42) repeating step S41) until all DS candidate points have been processed; obtaining the final result of the DS self-adaptive selection;
through the steps, the DS is selected in an InSAR time sequence analysis based on the land type in a self-adaptive mode.
2. The DS self-adaptive selection method in InSAR time sequence analysis based on land type as claimed in claim 1, wherein the step S1) of land type identification comprises the following steps:
s11) carrying out geometric registration on the N time sequence SAR images to obtain a registered time sequence SAR image; establishing a corresponding relation between SAR image pixel radar coordinates and geographic coordinates;
s12) determining the pixel position corresponding to the geographic coordinate of the SAR image pixel on the land type map by adopting a nearest neighbor method according to the geographic coordinate of the SAR image pixel in S11), thereby establishing the pixel corresponding relation between the land type map under the geographic coordinate system and the SAR image under the radar coordinate system, converting the land type map under the geographic coordinate system into the radar coordinate system, and finding out the land type attribute value (i.e. type value) corresponding to the SAR image pixel to realize the land type identification of the SAR image pixel.
3. The DS self-adaptive selection method in InSAR time sequence analysis based on land type as claimed in claim 1, characterized by, step S21) utilizing goodness-of-fit inspection to identify homogeneous pixels on the time sequence SAR image, specifically utilizing K-S inspection to identify homogeneous pixels on the time sequence SAR image;
before K-S detection, firstly determining the size m x n of a homogeneous pixel detection window, wherein m is the pixel number in the window orientation direction, and n is the pixel number in the distance direction;
then, judging the land type of the center pixel of the homogeneous pixel detection window: if the soil type is different from the soil type of other pixels in the window, the K-S inspection is not carried out on the soil type; otherwise, performing K-S test to determine homogeneous pixels.
4. The land type-based InSAR time sequence analysis DS self-adaptive selection method as claimed in claim 1, wherein the step S22) is to perform connectivity detection on the central pixels and the homogeneous pixels in the homogeneous pixel detection window, and the number of the pixels which are homogeneous and communicated with the central pixels and the window central pixels is used as the number of the homogeneous pixels in the window.
5. The method for self-adaptively selecting the DS in the InSAR temporal analysis based on the land type according to claim 1, wherein in step S23), the selection threshold of the DS candidate points is a positive integer and is not greater than the total number of pixels in the homogeneous pixel detection window, that is, the value is not greater than m × n; and when the number of the homogeneous pixels in the window is greater than or equal to the selection threshold of the DS candidate point, selecting the central pixel as the DS candidate point.
6. The method for adaptive selection of DS in InSAR time series analysis based on land type as claimed in claim 1, wherein in step S3), the processing method before adaptively calculating and determining the DS selection threshold comprises:
screening M image interference pairs from the N time sequence SAR images;
carrying out differential interference processing on the image interference pairs respectively to generate M differential interference images;
and setting the size of a coherency calculation window of the DS candidate points to be the same as that of a detection window of the homogeneous pixel by taking the DS candidate points as centers, wherein the size of the coherency calculation window of the DS candidate points is m x n.
7. The DS adaptive selection method in InSAR time sequence analysis based on land type as claimed in claim 6, characterized in that in step S3), specifically, for the ALOS-2 satellite strip mode SAR data, the time baseline threshold is set to 250-1500 days, and the space baseline threshold is set to 500-1500 m; for the Sentinel-1 satellite interference wide-amplitude mode SAR data, the time baseline threshold is set to be 60-200 days, and the space baseline threshold is set to be 100-250 m.
CN202010960574.9A 2020-09-14 2020-09-14 Land type-based DS self-adaptive selection method in InSAR time sequence analysis Active CN112130148B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010960574.9A CN112130148B (en) 2020-09-14 2020-09-14 Land type-based DS self-adaptive selection method in InSAR time sequence analysis

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010960574.9A CN112130148B (en) 2020-09-14 2020-09-14 Land type-based DS self-adaptive selection method in InSAR time sequence analysis

Publications (2)

Publication Number Publication Date
CN112130148A CN112130148A (en) 2020-12-25
CN112130148B true CN112130148B (en) 2021-12-28

Family

ID=73845668

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010960574.9A Active CN112130148B (en) 2020-09-14 2020-09-14 Land type-based DS self-adaptive selection method in InSAR time sequence analysis

Country Status (1)

Country Link
CN (1) CN112130148B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114814843B (en) * 2022-06-27 2022-10-04 之江实验室 Earth surface deformation inversion method and system based on CPU-GPU heterogeneous parallel

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101770026A (en) * 2009-01-07 2010-07-07 中国科学院电子学研究所 Method for estimating terrain by polarization interference of data of synthetic aperture radar and software thereof
CN108051810A (en) * 2017-12-01 2018-05-18 南京市测绘勘察研究院股份有限公司 A kind of InSAR distributed diffusions body phase optimization method
CN108387899A (en) * 2018-04-17 2018-08-10 南京师范大学 Ground control point automatically selecting method in synthetic aperture radar interferometry
CN110058237A (en) * 2019-05-22 2019-07-26 中南大学 InSAR point Yun Ronghe and three-dimensional deformation monitoring method towards High-resolution SAR Images

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101770026A (en) * 2009-01-07 2010-07-07 中国科学院电子学研究所 Method for estimating terrain by polarization interference of data of synthetic aperture radar and software thereof
CN108051810A (en) * 2017-12-01 2018-05-18 南京市测绘勘察研究院股份有限公司 A kind of InSAR distributed diffusions body phase optimization method
CN108387899A (en) * 2018-04-17 2018-08-10 南京师范大学 Ground control point automatically selecting method in synthetic aperture radar interferometry
CN110058237A (en) * 2019-05-22 2019-07-26 中南大学 InSAR point Yun Ronghe and three-dimensional deformation monitoring method towards High-resolution SAR Images

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
一种适用于时序InSAR的同质像元加权干涉相位滤波方法;黄俊松 等;《北京大学学报（自然科学版）》;20181130;第54卷(第6期);全文 *
时序InSAR同质样本选取算法研究;蒋弥 等;《地球物理学报》;20181231;第61卷(第12期);全文 *

Also Published As

Publication number Publication date
CN112130148A (en) 2020-12-25

Similar Documents

Publication Publication Date Title
CN110443836A (en) A kind of point cloud data autoegistration method and device based on plane characteristic
CN113960595A (en) Surface deformation monitoring method and system
CN110837006B (en) Satellite lightning detection evaluation method based on satellite-ground synchronous observation and comparison
CN112965146A (en) Quantitative precipitation estimation method combining meteorological radar and rainfall barrel observation data
CN109583284A (en) Urban skyscraper object height extracting method and device based on High Resolution SAR Images
CN108230375A (en) Visible images and SAR image registration method based on structural similarity fast robust
CN112130148B (en) Land type-based DS self-adaptive selection method in InSAR time sequence analysis
CN111522977A (en) Karst abandoned mine ecological restoration remote sensing detection method
CN113281749A (en) Time sequence InSAR high-coherence point selection method considering homogeneity
CN115984778A (en) Multi-feature optimization based method for rapidly and dynamically monitoring Sentinel-1 data in flood
CN106940782A (en) High score SAR based on variogram increases construction land newly and extracts software
CN116012364A (en) SAR image change detection method and device
CN114387408A (en) Method and device for generating digital elevation model and computer readable storage medium
CN112989940B (en) Raft culture area extraction method based on high-resolution third satellite SAR image
CN106932777A (en) Interfering synthetic aperture radar based on temperature baselines is to optimum option method
CN116908853B (en) High coherence point selection method, device and equipment
CN107067384B (en) A kind of loess plateau terraced fields extracting method based on ray method
CN113687353A (en) DS target phase optimization method based on homogeneous pixel time sequence phase matrix decomposition
CN114266899A (en) Image target parallel detection method based on multi-core DSP
CN116740579B (en) Intelligent collection method for territorial space planning data
CN111915570A (en) Atmospheric delay estimation method based on back propagation neural network
CN113008202B (en) Ground settlement monitoring method integrating different synthetic aperture radar interferometry
CN109387872A (en) Surface-related multiple prediction technique
CN111239734B (en) Extraction method suitable for deep loess stable surface scatterers
CN110874833A (en) SAR image change detection method based on hypergraph matching

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant